The current proposal is aimed at developing a novel neurobiological marker based on patterns of whole-brain functional connectivity during implicit and subliminal threat-related facial emotion perception to aid the diagnosis and assessment of treatment response in social anxiety (SAD) and panic disorder (PD). The following Specific Aims are proposed: Specific Aim 1). Develop, apply and validate an approach to estimate the large-scale functional networks of implicit and subliminal threat-related and ambiguous facial emotion processing in a healthy control subject pool. We will perform this by first testing whether large-scale partial correlations measured during a particular condition (i.e. presentation of visible or invisible fearful and neutral faces) can be used as features to predict the stimulus type presented during the condition (Sub-aim 1a). We will then test whether the most informative features for predicting fearful vs. neutral face conditions are connections among primarily limbic and paralimbic nodes, and that the most informative features in predicting neutral vs. resting states conditions are connections among primarily attention, motor and sensory related regions (Sub-aim 1b). Specific Aim 2). Determine whether the approach developed in SA1 may have potential clinical application towards the diagnosis and treatment of SAD and PD. We will perform this by testing whether the above functional networks are informative brain signatures that can be used to discriminate between healthy controls, SAD and PD subjects (Sub-aim 2a). We will also test whether the approach can be used to determine the particular sub-networks that are aberrant in SAD and PD (Sub-aim 2b).